in high‐resolution images even when reformatted. CB‐CT is not quantitative and does not produce a measurement of HU.
Image processing occurs through mathematical manipulation of the density data and has a profound impact on the appearance and clinical utility of an image. Unprocessed or raw CT data are typically not used in diagnostic imaging and may not even be stored by the acquisition device or picture archiving and communication system (PACS). Most CT scanners have several processing algorithms that allow the operator to choose the degree and type of processing at the time of acquisition. The methods of processing evolve but in general will include bone, sharp or edge‐enhanced algorithms along with soft tissue or smoothing algorithms. Complete examination of an anatomic region should include both so that all tissues can be evaluated. A sharpening algorithm will produce very pleasing diagnostic images of bone and fractures but will enhance artefacts such as high‐density edge gradient artefacts causing streaking through regional soft tissues.
Image display is flexible. The end user is able to selectively manipulate the image to emphasize structures of different density. Window width refers to the range of HU over which the greyscale is applied, and window level refers to the centre point of the window. In order to fully evaluate a region, both window level and width require manipulation.
CT produces excellent bone images due to the inherent high subject contrast when using tissue density/X‐ray attenuation (400–2000 HU). It is particularly good for imaging fractures due to the combination of high inherent contrast between intact and disrupted bone and high spatial resolution that permits identification of very small areas of disruption. In principle, soft tissues have less inherent contrast and are imaged less well. Modern scanners, capable of high tube output, produce very good soft tissue image quality, although when immediately adjacent to a high‐density tissue, such as cortical bone, this can be more problematic.
Artefacts
CT, like all imaging modalities, has its own complement of artefacts. These are defined as a discrepancy between the CT number or HU in the reconstructed image and the actual attenuation coefficient of the object. Non‐conventional use of CT technologies, such as standing CT, results in an additional gamut of artefacts that must be understood and evaluated for what they are.
Partial volume averaging results in the incorrect assignment of an HU value when the values of two structures are averaged in one voxel. This is problematic in fracture identification if the fracture is non‐ or minimally displaced and/or running obliquely through the scan plane but can be mitigated by reformatting the images into multiple different planes.
High‐density edge gradient or beam hardening occurs when a very dense subject is present in the scan plane, attenuating the low‐energy portion of the polychromatic photon beam and resulting in a preponderance of higher energy X‐rays. This results in dark bands or streaks either between two high‐density structures (e.g. petrous temporal bone) or around the margins of a high‐density structure such as a metallic implant. Beam hardening can be difficult to avoid in equine patients. Most CT scanners have beam hardening reduction software that may or may not be available to the operator. Photon starvation is caused by beam hardening between two dense objects. This is of particular importance in horses when two limbs are placed through the gantry at the same time. Even if the operator reduces the field of view to include only one limb, the effect of the pair will be visible in the images.
Motion produces image blurring or mismapping of anatomy. These can have negative impacts on the identification of fragments if the blurring causes margins to become inconspicuous or in fracture evaluation when a hypoattenuating area such as fracture gap can be mismapped to a different region.
Photon starvation is seen in areas of high attenuation, particularly associated with metal implants. Insufficient photons reach the detector, and during reconstruction noise is greatly magnified in these areas creating streaks in the image.
Clinical Indications
In anatomically accessible areas, CT has the potential to provide additional and useful information for the identification and characterization of all fractures, whether they are managed conservatively or with surgical intervention. The benefits must be weighed against the potential risks associated with acquisition such as general anaesthesia and moving the horse to or through the scanner.
CT is considered the gold standard for fracture diagnosis and evaluation of three‐dimensional configuration. Complex, comminuted, articular fractures, small, minimally displaced fractures of long bones or simple fractures in complicated anatomic regions are best evaluated with cross‐sectional CT imaging with or without 3D or surface rendering. In humans and horses, CT has been shown to be more sensitive than radiographs for identifying fractures and recognizing comminution [117–121].
The three‐dimensional nature of CT has proved integral to presurgical planning and has been reported for the central tarsal bone [122], distal phalanx [123, 124], navicular bone [124] and proximal phalanx [125]. This is also the case in the authors experience for third carpal bone fractures (Figure 5.11); further applications are documented throughout the book. It has been repeatedly shown to give better spatial information and thus recognition of fracture configuration and complexity and the structure of affected bones and fragments [126]. In addition, areas with complex anatomy or shape, such as the distal phalanx, where dimensions vary according to orientation, and cases with multifocal pathology are only adequately assessed by CT [123, 126, 127].
Osseous trauma of the skull is better evaluated with CT than plain radiographs with respect to identification [128], classification and surgical planning [129], although small fractures maybe missed if inappropriate window parameters are chosen [130] (Chapter 36). The basics of acquisition, i.e. thin slice thickness, and appropriate reading, i.e. bone algorithms, are essential [131]. CT can also differentiate between structures that radiographically mimic fractures such as suture lines or overlapping sinuses.
Figure 5.11 Evaluation and surgical planning of two‐third carpal bone fractures. (a) Dorsal 35° proximal–dorsodistal oblique radiograph demonstrating a parasagittal plane fracture of the radial facet and corresponding dorsal plane reformatted CT image revealing the fracture line to extend from the middle carpal joint to the distal subchondral bone plate. A lag screw was therefore placed in a central position in the bone. (b) Flexed dorsal 35° proximal–dorsodistal oblique radiograph demonstrating a dorsal plane fracture of the radial facet and corresponding sagittal plane reformatted CT image demonstrating the fracture to be located in the proximal third of the bone. The surgical implant was therefore placed proximally in the bone at the mid‐point of the fracture.
Small, portable CT machines can be used during surgical procedures. CT‐assisted surgery of navicular bone and distal phalangeal fractures has increased surgical accuracy and reduced surgery time. Barium paste as markers for orientation applied to the hoof wall [124], and surgical skin staples [122] have been used as surface locators.
Limitations
CT is an excellent determinant of bone morphology but does not provide information about biological activity. This can be inferred by interpretation of the complement of morphological changes but does not reflect the level of activity as seen in nuclear medicine studies (scintigraphy or positron emission tomography [PET] scanning) or provide a visual map of intra‐osseous fluid accumulation as shown by fluid‐sensitive
